Component supply device and machine learning device

A component supply device that supplies a plurality of components placed on a tray and determines operation parameters for a vibration operation of the tray through which the arrangement and the postures of the components placed on the tray are changed, the component supply device comprising:

• a machine learning device that learns the operation parameters for the vibration operation, wherein
• the machine learning device includes
• a state observation unit that observes vibration operation parameter data representing the operation parameters for the vibration operation and component arrangement data representing the arrangement and the postures of the components on the tray, as state variables representing the current state of the environment,
• a determination data acquisition unit that acquires determination data representing a suitability determination result of the vibration operation, the suitability determination result representing efficiency in supply of the components, and
• a learning unit that learns the operation parameters for the vibration operation and the component arrangement data while associating the operation parameters for the vibration operation with the component arrangement data, by using the state variables and the determination data,
• wherein the learning unit includes:
• a reward calculation unit that obtains a reward related to the suitability determination result;
• and
• a value function update unit that updates a function representing the values of the operation parameters for the vibration operation with respect to the arrangement and the postures of the components on the tray by using the reward.

A component supply device that supplies a plurality of components placed on a tray and determines operation parameters for a vibration operation of the tray through which the arrangement and the postures of the components placed on the tray are changed, the component supply device comprising:

• a machine learning device that learns the operation parameters for the vibration operation, wherein
• the machine learning device includes
• a state observation unit that observes vibration operation parameter data representing the operation parameters for the vibration operation and component arrangement data representing the arrangement and the postures of the components on the tray, as state variables representing the current state of the environment,
• a determination data acquisition unit that acquires determination data representing a suitability determination result of the vibration operation, the suitability determination result representing efficiency in supply of the components, and
• a learning unit that learns the operation parameters for the vibration operation and the component arrangement data while associating the operation parameters for the vibration operation with the component arrangement data, by using the state variables and the determination data,
• wherein the learning unit learns the operation parameters for the vibration operation in each of a plurality of component supply devices by using the state variables and the determination data, the state variables and the determination data being obtained for each of the plurality of component supply devices.

A machine learning device for learning-operation parameters for a vibration operation of a tray through which an arrangement and a posture of each of a plurality of components scattered and placed on the tray are changed in a component supply device for supplying the plurality of components placed on the tray, the machine learning device comprising:

• a processor configured to:
• Observe state variables representing a current state of an environment, the state variables including:
• vibration operation parameter data representing the operation parameters for the vibration operation, and component arrangement data representing the arrangement and the posture of each of the plurality of components on the tray;
• acquire determination data representing a suitability determination result of the vibration operation, the suitability determination result representing efficiency in supply of the plurality of components, and by using the state variables and the determination data, learn the operation parameters for the vibration operation and the component arrangement data while associating the operation parameters for the vibration operation with the component arrangement data, wherein the operation parameters include at least one of time of vibration and a phase difference of the vibration operation.